swift-sqlite-data

Pass

Audited by Gen Agent Trust Hub on Mar 28, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill documents patterns for ingesting user-supplied data into database queries, which is a standard surface for indirect prompt injection.
  • Ingestion points: Untrusted user input is ingested via the searchText and searchTokens parameters in the full-text search (FTS5) logic documented in references/advanced-queries.md.
  • Boundary markers: The skill demonstrates the use of type-safe query builders and property wrappers (@FetchAll, @FetchOne) to scope and delimit queries.
  • Capability inventory: The skill provides instructions for database writes (database.write), schema migrations, and remote data synchronization using Apple's CloudKit (SyncEngine).
  • Sanitization: The provided example for buildFTS5Query uses simple string concatenation for building search components; developers must ensure the underlying library handles or they manually escape FTS5 control characters.
  • [SAFE]: The skill consists entirely of documentation, guidelines, and code snippets for developer use, with no signs of obfuscation, malicious persistence, or credential harvesting.
  • [EXTERNAL_DOWNLOADS]: The skill describes synchronization with CloudKit, which is a well-known service and a legitimate part of the library's intended functionality for data persistence and sharing.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 28, 2026, 12:37 PM
Security Audit — agent-trust-hub — swift-sqlite-data